Table partitioning is a mechanism to segregate a large table into smaller, more manageable parts without creating a separate table for each part. A partitioned table physically stores data in groups of rows. These groups of rows are called partitions. You can access and maintain each partition separately.
DynamoDB also manages data in partitions. DynamoDB uses SSDs for storing data and automatically replicates data across multiple AZs in an AWS region. DynamoDB automatically manages partitions; you, as a consumer, do not need to manage the partitions.
When creating a table, DynamoDB allocates a sufficient number of partitions to the new table so that it can handle any provisioned throughput needs. However, DynamoDB can allocate additional partitions to a table in certain situations. The following are the scenarios when DynamoDB allocates additional...